Touch Device Control System Using CNN for Object Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Touch devices face challenges in accurately determining the category of objects in contact or their status, leading to inefficiencies in performing corresponding operations and recalibration, as existing systems lack precise recognition mechanisms.
Innovation Solution
A control system incorporating a convolutional neural network (CNN) processes sensing images from touch sensors to generate feature and identification information, enabling precise determination of object categories and device status, utilizing a sensing circuit, processor, and host to perform object segmentation and recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional sensing mechanisms are used to determine object category or device status, then the system structure remains simple, but the recognition accuracy and precision are insufficient
Solution Approach 1:
The patent replaces traditional mechanical sensing and rule-based recognition systems with a convolutional neural network (CNN) based intelligent recognition system. The CNN processes sensing images to automatically identify object categories and device statuses, achieving high recognition accuracy without requiring complex manual feature engineering or multiple specialized sensors.
Solution Approach 2:
The patent transforms the sensing data from raw sensing values into sensing images, then processes these images through CNN to extract feature information and generate identification information. This parameter transformation approach enables the system to achieve high recognition precision by changing the data representation form rather than increasing hardware complexity.
2Measurement precision
If complex recognition mechanisms are implemented to improve object category determination, then recognition accuracy improves, but processing time and system complexity increase
Solution Approach 1:
The patent performs preliminary processing by converting sensing values into sensing images before CNN processing. This pre-processing step organizes the data in a format optimized for neural network processing, enabling faster and more accurate object category recognition while reducing the computational burden during the actual recognition phase.
Solution Approach 2:
The patent uses CNN-based intelligent processing to replace traditional complex recognition algorithms. The CNN architecture is specifically designed to efficiently process image data and extract relevant features, achieving high recognition accuracy with optimized processing time compared to conventional approaches.
3Productivity
If traditional sensing and determination methods are used, then the system requires less computational power, but the ability to perform precise operations and recalibration is limited
Solution Approach 1:
The patent implements a universal CNN-based recognition system that can handle multiple tasks including object category identification, device status determination, and anomaly detection. This multi-functional approach improves operational efficiency by using a single intelligent system rather than multiple specialized mechanisms, optimizing the balance between computational energy consumption and productivity.
Data Source
AI summary
A control system applicable to a touch device is provided. The touch device includes a touch sensor. The control system comprises a sensing circuit, configured to sense the touch sensor and generate a plurality of sensing values; a processor, configured to generate a sensing image according to the plurality of sensing values; and a convolutional neural network, configured to process the sensing image to generate feature information and generate identification information according to the feature information for determining a status of the touch sensor.